Predictive Modeling of Repeat Hotel Visits

In recent years, the number of studies on the subject of customer satisfaction and retention in the hospitality industry has grown. The primary reason is advances in technology, especially in data analytics and machine learning, which have made it possible to use large amounts of guest data to gain insight into guest behavior and preferences. In this study, a robust predictive model is aimed to be developed to forecast the likelihood of repeat visits to hotels. A comprehensive dataset comprising 83590 instances over three years (2015-2018, Lisbon), encompassing personal, behavioral, demographic, and geographical information, is leveraged for an in-depth analysis to identify key factors influencing repeat visits. Research objectives include pinpointing critical factors shaping repeat visits, assessing their relative importance in guests' overall satisfaction levels, and evaluating the efficacy of various predictive models. By scrutinizing customer behavior through logistic regression and advanced Random Forest models, insights into guest preferences and tendencies are unearthed. The findings underscored the significance of lead time, lodging revenue, and customer interactions, such as canceled bookings, in predicting repeat visits. Notably, extended booking lead times were found to deter customer returns, while higher lodging revenues fostered greater guest loyalty. Conversely, higher nightly rates posed a deterrent to repeat visits, highlighting the need for strategic pricing adjustments. Transitioning to Random Forest models substantially augmented predictive capabilities, enabling nuanced exploration of complex data relationships and yielding near- flawless classification accuracy. This underscored the importance of continual model refinement based on empirical data. By leveraging predictive modeling, hotels can tailor their services to meet guest expectations effectively, driving long-term success in a competitive market landscape. Keywords: Predictive modeling, Repeat hotel visits, Customer retention, Logistic regression, Random Forest